Ordered data point cloud obtained by laser-scanning, image processing or contact digitizing techniques is widely used for reconstructing a cross section profile in the field of reverse engineering. In this research, a comprehensive algorithm for reconstruction of 2D profile is developed based on NURBS parametric curve theory. In this regard, the line, arc and B-spline segments are firstly extracted from the set of ordered data points using a developed segment detection algorithm. Then, the final profile is reconstructed using the newly proposed algorithm of constrained local fitting which approximates these segments by NURBS curves with appropriate geometric continuity conditions. The validity, capability and functionality of the developed method are investigated by some practical case studies. Results show that developed algorithm can be integrated and improve the functionality of exiting reverse engineering systems.
